The Life of a Data Byte

> UTF-8 is the standard for representing characters as eight bits, allowing every code-point between 0-127 to be stored in a single byte...

Hmm... this makes it sound like UTF-8 uses one byte per codepoint and can't encode codepoints beyond 127, when it actually is a variable-length encoding where one codepoint is encoded into 8 to 48 bits (1 to 6 bytes, although in practice 1 to 4 bytes are sufficient to encode all currently defined UNICODE codepoints).

The same is true for UTF-16, it's not "16 bits per codepoint", but variable length with either 16- or 32-bits per codepoint. Only UTF-32 is a fixed-width encoding.

Basically: all UTF encodings are able to encode all UNICODE codepoints.

Cool article. Another future method is the use of DNA (e.g. [1], [2]). Not good for speed, but with more development it could become a great solution for large-volume, low-access archival data.

[1] https://www.microsoft.com/en-us/research/project/dna-storage... [2] https://www.catalogdna.com

> Optane claims performance 1,000 faster than NAND SSDs with 1,000 times the performance, while being four to five times the price of flash memory. Optane is proof that storage class memory is not just experimental.

This was a very bold claim from Intel. That was in 2015. It’s 2020 and I haven’t seen any benchmarks that meet this. And it’s not clear the roadmap is going to scale to that figure. Right now it’s still not quite good enough to be a main memory (but as a caching tier between main memory and disk).

So sad that the mack daddy of removable storage in the late 90's isn't covered. Jaz disks were freaking boss for all of the 5 years they existed.

Small nit, I'm pretty sure 1980 5¼'' floppies where 360 Kbytes and not the later 1.2 MBytes.

The "rope memory" used on Apollo seems like such a crazy and cool idea. The article has one dead link, and the reference only has a snippet more information. I'm about to go down a google-hole on it, but can anyone point me to their favorite rope-memory resource?